Most Popular SCADA Systems and Where Each One Fits

Ask ten engineers to name the most popular SCADA systems and you get ten different lists. That is not because anyone is wrong. It is because “popular” depends on which room you are standing in.

Walk into an automotive plant and you will see FactoryTalk View SE on the wall. Walk into a distribution control center two kilometers away and you will see MicroSCADA X or Survalent. Both are SCADA. Neither vendor competes for the other’s projects.

So this page splits the field by where the software actually gets deployed, not by a made-up ranking. Each entry covers what the platform is good at, how it licenses, and which protocols it speaks natively.

The short version

Tier 1: General-purpose plant platforms

These are the products that show up in factories, water plants, and PV farms. They talk to whatever PLC is in the panel and they do not assume a specific vendor’s hardware.

Ignition (Inductive Automation)

Ignition is the platform that changed how the industry thinks about SCADA pricing. You buy a server license. Client count is unlimited. Tag count is unlimited. That single decision is why it spread so fast among integrators who were tired of quoting tag packs.

Version 8.3 arrived in September 2025 as a long-term support release, with updates promised across a five-year window. It brought a built-in REST API, Git-based version control for projects, a rebuilt Siemens driver with symbolic browsing, and much simpler containerized deployment. Point releases have kept coming since, adding OPC UA Global Discovery Server support and file-based secret providers.

The platform is Java and Kotlin under the hood, so it runs on Windows, Linux, and macOS. That matters more than it sounds. A Linux gateway in a Docker container is a very different operational story than a Windows server that needs patch windows.

Modules do the real work. Perspective builds responsive web and mobile screens. Vision handles traditional desktop clients. There is a built-in OPC UA server, a SQL bridge, a reporting engine, and MQTT modules from Cirrus Link for Sparkplug B publishing.

Native protocols: Modbus RTU and TCP, Allen-Bradley (Logix, PLC-5, SLC), Siemens S7, DNP3, BACnet, OPC UA, MQTT with Sparkplug B.

Where it struggles: IEC 61850 needs a third-party module or a protocol gateway in front of it. On a substation project where the SCADA layer has to sit directly on the station bus, this is not the shortest path.

SIMATIC WinCC (Siemens)

WinCC is not one product. This trips up a lot of people, so here is the split:

• WinCC Basic, Comfort, Advanced — panel and single-station HMI, configured inside TIA Portal.
• WinCC Professional — the SCADA-scale option inside TIA Portal, sometimes called RT Professional.
• WinCC Unified — the newer HTML5-native line, also in TIA Portal, currently shipping with V20.
• WinCC V7 — the older standalone product, still installed everywhere.
• WinCC Open Architecture (WinCC OA) — a completely different codebase, originally PVSS from ETM.

The pull is integration. When your PLC program and your HMI live in the same engineering project, you tag-browse straight from the controller. No OPC layer, no mapping spreadsheet, no drift between the two. In a S7-1500 plant that saves real hours.

WinCC OA deserves separate mention. It is event-driven, runs distributed across many servers, and scales to millions of data points. CERN runs its accelerator control on it. Airports, tunnels, and metro lines use it too. It is not the product you pick for a small skid.

Native protocols: S7 communication, PROFINET, PROFIBUS, OPC UA, Modbus TCP. WinCC OA adds IEC 60870-5-104, IEC 61850, and DNP3 drivers.

Where it struggles: licensing is priced by runtime power tags, and the tiers step up hard. Third-party PLC support is workable but never as smooth as native S7.

FactoryTalk View SE (Rockwell Automation)

Site Edition is the distributed version, with HMI servers, data servers, and clients spread across a network. It is the default answer in any plant standardized on Allen-Bradley.

FactoryTalk Linx handles the communication layer, browsing ControlLogix and CompactLogix tags directly over EtherNet/IP. Alarms come from the controller itself through FactoryTalk Alarms and Events, which means alarm state survives an HMI restart.

Rockwell also ships FactoryTalk Optix now, a newer cross-platform product built around OPC UA and web clients. It targets a different buyer, but it signals where the portfolio is heading.

Native protocols: EtherNet/IP and CIP, ControlNet, DeviceNet, OPC UA, Modbus TCP through gateway drivers.

Where it struggles: outside the Rockwell ecosystem, the value drops fast. Non-CIP devices go through OPC or third-party drivers, and you lose the tight integration that justified the price.

AVEVA System Platform and InTouch HMI

InTouch has been in service since the late 1980s under the Wonderware name. That history is the product’s biggest asset and biggest drag at the same time.

The asset: application longevity. An InTouch application built fifteen years ago still opens in the current release. In a refinery or a paper mill where the last rebuild was in 2009, that matters enormously.

The drag: architectural layers. System Platform sits above InTouch and provides the object model, originally called ArchestrA. You build templates for a pump, instantiate two hundred of them, and change all two hundred by editing the template. Powerful, but the learning curve is steep and the deployment model has a lot of moving parts.

The 2023 R2 line is current, with SP1 shipping since. Recent work has focused on SVG import for graphics, user-defined types, and OPC UA tag creation in bulk.

The wider portfolio matters here too. AVEVA Historian, AVEVA PI System from the OSIsoft acquisition, and Plant SCADA (the old Citect product) all sit under the same roof.

Native protocols: Modbus, OPC UA and DA, DNP3, ABB and Siemens drivers, plus a large third-party DAServer catalog.

Where it struggles: the licensing and subscription model has moved around several times. Budget planning is harder than it should be.

Proficy iFIX and CIMPLICITY (GE Vernova)

Two products, one portfolio, different roots. iFIX came from Intellution. CIMPLICITY was GE’s own. Both are still actively developed and both use year-based release naming, so you will see iFIX 2026 and CIMPLICITY 2026 rather than version numbers.

iFIX leans toward process visualization with a strong graphics engine and model-based navigation. CIMPLICITY leans toward enterprise-scale event management and is common in automotive and in power generation.

Recent releases have pushed HTML5 clients through Proficy Operations Hub, centralized certificate management in Configuration Hub, Python scripting extensions, and native MQTT publishing.

Native protocols: Modbus, OPC UA and DA, MQTT, Allen-Bradley and Siemens drivers, IEC 60870-5-104 and DNP3 through the utility driver set.

Where it struggles: the portfolio has been reorganized repeatedly through the GE Digital to GE Vernova transition. Documentation and support paths can be fragmented.

zenon (COPA-DATA)

A smaller Austrian vendor with an outsized presence in two niches: energy and regulated pharma.

The energy edition ships genuine IEC 61850 client and server support, IEC 60870-5-101, -103, and -104, DNP3, and Modbus, all in the base product rather than as paid add-ons. For a substation gateway or a hybrid plant where you need to bridge protection relays to a corporate historian, that protocol depth is the whole selling point.

On the pharma side, zenon carries audit trail, electronic signature, and validation support aimed at GMP environments.

Where it struggles: smaller integrator base outside Europe. Finding local support in some regions takes effort.

ICONICS Suite (Mitsubishi Electric)

GENESIS64 was one of the first SCADA products built natively on 64-bit Windows and OPC UA rather than OPC DA. Mitsubishi acquired ICONICS in 2019 and the product now sits inside their portfolio.

Hyper Historian is the strong piece, with high-speed collection and swinging-door compression. AnalytiX adds fault detection. The building automation market is where ICONICS has the deepest install base, especially for large campus and data center energy monitoring.

VTScada (Trihedral)

A single integrated product rather than a stack of modules. Historian, alarm system, reporting, and redundancy are all built in.

Municipal water and wastewater is the core market, particularly in North America. The redundancy model is genuinely good: any workstation can act as a server, and the system self-heals when a node drops. For a utility running twenty remote lift stations over radio, that resilience is worth more than a fancy graphics engine.

Licensing is per tag and perpetual, which some utilities prefer over subscription.

Tier 2: Utility and grid platforms

Different world. These systems manage networks, not processes. They model topology, run state estimation, handle outage management, and hold NERC CIP or equivalent compliance evidence.

MicroSCADA X SYS600 (Hitachi Energy)

Built for substation automation and network control, with decades of history and a very large installed base across transmission and distribution utilities.

One naming point worth clearing up, because it still causes confusion on tender documents and CVs. MicroSCADA was an ABB product for most of its life. It moved to Hitachi with the acquisition of the ABB power grids business, and now sits in the Hitachi Energy substation automation, protection, and control portfolio alongside the RTU500 series and the Relion relays. Manuals from the 10.x line already carry Hitachi copyright. If you see it listed under ABB anywhere, the source is out of date.

SYS600 handles station-level SCADA, gateway functions, and network control. SCIL is the scripting language, which is unusual but powerful once learned. The Historian module covers long-term process data. DMS600 adds distribution management on top, including fault isolation and service restoration sequences that SYS600 then executes. SDM600 handles disturbance record collection, central account management, and relay firmware tracking over IEC 61850.

SYS600C is the compact hardware variant for smaller substations and gateway duty.

Protocols: IEC 61850 MMS as client and server, IEC 60870-5-101 and -104, IEC 60870-5-103, DNP3, Modbus, and ICCP/TASE.2 for control center exchange.

Network Manager (Hitachi Energy)

The same vendor, a tier above. Network Manager is the transmission-scale SCADA/EMS/ADMS platform: state estimation, contingency analysis, load flow, and generation dispatch. This is control center software, not station software. A utility running MicroSCADA X in its substations and Network Manager in its control room is a normal architecture, with ICCP linking the two.

Spectrum Power (Siemens)

Siemens’ equivalent at the same tier, covering transmission SCADA/EMS and a distribution ADMS variant. Common in European TSO and DSO control rooms.

monarch (AspenTech, an Emerson company)

Open Systems International built monarch as a single platform covering SCADA, EMS, DMS, and OMS on a common data model. Emerson acquired OSI, then folded it into AspenTech. The platform kept its momentum through the ownership changes and is a regular finalist in North American utility tenders.

SurvalentONE (Survalent)

The mid-market answer. Municipal utilities and cooperatives that need real ADMS capability without a transmission-scale budget. Integrated SCADA, OMS, and DMS with no artificial point limits.

EcoStruxure Geo SCADA Expert (Schneider Electric)

Formerly ClearSCADA. Purpose-built for telemetry over unreliable links, which makes it strong in water networks, oil and gas gathering, and pipeline monitoring.

The DNP3 implementation is one of the better ones in the market, with proper handling of unsolicited responses, event buffering, and time synchronization across poor-quality links. The database is object-oriented, and the historian is embedded rather than bolted on.

ViewX is the thick client, WebX the browser client.

e-terraplatform (GE Vernova)

The EMS side of the GE Vernova grid software portfolio, used by large system operators for real-time network operations and market functions.

Tier 3: Free and open-source options

Worth knowing about, but be honest about what you are signing up for.

Rapid SCADA — open source, actively maintained, modular. Runs on Windows and Linux. Modbus, DNP3, OPC UA, and MQTT drivers exist. Reasonable for small monitoring jobs and for learning.

FUXA — a Node.js web-based SCADA. Modern browser interface, Modbus, OPC UA, MQTT, S7, and BACnet support. Lightweight enough to run on a Raspberry Pi.

ScadaLTS — a web-based fork of the old Mango/ScadaBR lineage. Java-based, decent driver set.

Ignition Maker Edition — not open source, but free for non-commercial use. The full platform with the commercial license removed. The best way to learn Ignition properly without a trial clock running.

The catch with all open-source SCADA is not the software. It is the absence of a vendor to call at 3 a.m. when the historian stops writing and the plant manager is standing behind you. For a home lab, a training rig, or a low-consequence monitoring job, they are fine choices. For anything with safety or revenue attached, the support contract is the product.

What the software actually runs on

Almost nobody asks this question until the quote comes back and there is a line item for a server nobody budgeted. Worth covering, because the answer differs sharply by tier.

Most SCADA vendors do not build servers. They publish a sizing spec and certify third-party platforms. AVEVA, GE Vernova, Inductive Automation, and COPA-DATA all work this way. You buy Dell or HPE, you match the datasheet, you are supported.

Three vendors are exceptions.

Siemens SIMATIC Rack PC

Siemens sells industrial computers in three classes, and the naming does not make the split obvious.

Server class is the RS-828A: a 2U rack machine with two AMD EPYC Embedded 9004 processors, up to 96 cores, DDR5 across 24 DIMM slots, dual 10 GbE on board, and PCIe Gen 5 expansion. Rated for 45 °C ambient and built to keep performance under vibration. Siemens positions it explicitly for virtualized WinCC servers, virtual HMI clients, and virtual PLCs on one box.

Workstation class covers the IPC647E, IPC847E, and IPC1047E, plus the newer RW-543B and RW-545A. The 847E is the traditional WinCC server box: 4U, RAID 1, industrial power supply, and a lot of PCI slots for legacy communication cards. Client class is the RC-543B and RC-545A.

RAID 1 or 5 and redundant power supplies are available across the rack line. That, plus a documented ten-year availability window, is what you are paying the premium for over a generic server.

Rockwell Automation

Rockwell’s industrial computers are client-class hardware, not servers. The line came from the ASEM acquisition and is now branded ASEM 6300, covering box PCs and panel PCs. It was sold as VersaView 6300 for several years, so both names appear in the field. The older VersaView 5000-series computers and thin clients are discontinued.

For the server side, Rockwell sells the Industrial Data Center: a pre-assembled rack built on Cisco UCS servers with VMware vSphere, shipped factory-configured with on-site commissioning and a single support number. Two sizes, with two or three UCS nodes. The point is not the hardware, which is standard. The point is that one vendor owns the whole stack when something breaks at 2 a.m.

Hitachi Energy SYS600C

A compact rack-mounted station computer built for the substation environment rather than a server room. It runs MicroSCADA X as a gateway, a station HMI, or a communication server, and ships with the software already loaded. Redundant power supplies are available as an option, and the unit carries a large port count for serial and Ethernet communication.

This is the category worth understanding if you work in substations. Computers and switches installed in a switchyard need IEC 61850-3 and IEEE 1613 compliance: wide temperature range, conformal coating, and immunity to the electrical noise a substation generates. A standard rack server in a control building is fine. The same server in a relay house is not.

The fault-tolerant option

Stratus ftServer, now a Penguin Solutions brand, sits in its own category. Duplicated hardware inside a single chassis running in lockstep on custom FPGA firmware. Both halves execute the same instruction at the same time. A component failure causes no interruption and no data loss, and the operating system sees one machine.

The economics are the real argument. One logical server means one Windows license, one SCADA license, one database license, one antivirus subscription. A redundant pair of physical servers means two of everything. On a remote site with no local IT, the math often lands in favor of lockstep.

AVEVA certifies System Platform on it. Schneider packages it inside the EcoStruxure Micro Data Center enclosure with a UPS and environmental monitoring.

One constraint to design around: running a FactoryTalk View SE server inside a VM on ftServer is documented by Rockwell as breaking CPU lockstep redundancy status. Bare-metal deployment is the supported path there. Confirm the current knowledgebase article before you commit an architecture either way.

Everything else

Control rooms and plant server rooms run standard enterprise iron. Dell PowerEdge and HPE ProLiant 2U machines are the default. Dual power supplies on separate feeds, ECC memory, RAID, out-of-band management. Utility EMS and ADMS platforms typically sit on Red Hat Enterprise Linux; plant SCADA sits on Windows Server.

Virtualization is now the normal case rather than the exception. vSphere held the default position for years, and the licensing changes since the Broadcom acquisition have pushed a lot of sites to evaluate Hyper-V, Nutanix, and Proxmox. Check the vendor support statement before you assume. Hypervisor support is stated per product version, and serial communication hardware and anything with hard timing requirements do not always virtualize cleanly.

Sizing, in practice

Three rules cover most of it.

The historian is the bottleneck, not the SCADA runtime. Tag count drives memory; scan rate and retention drive disk I/O. Put the historian and the SQL database on NVMe, and give them separate volumes from the operating system.

Network interfaces stay separated. One NIC on the process network, one on the corporate side, and no bridging between them. This is a segmentation requirement long before it is a performance one.

Size for the system you will have in five years. Point counts grow every time someone adds a feeder, a string inverter, or a metering panel. Rebuilding a production SCADA server because it ran out of RAM is a bad weekend.

PRP and HSR at the server

Network redundancy is a separate question from server redundancy, and it gets skipped in most hardware discussions. Under IEC 62439-3, PRP and HSR both give zero recovery time by sending duplicate frames over independent paths. The receiver takes the first copy and drops the second. Nothing fails over, because nothing was ever waiting to fail over. The full comparison is covered in PRP vs HSR.

For a SCADA server, there are only two ways to get there.

As a doubly attached node. The server itself runs the PRP stack across two physical adapters and presents one virtual interface to the operating system. Siemens supports this on SIMATIC IPCs with SOFTNET-IE RNA, a licensed software package that builds the virtual adapter over two identical NICs. The requirements are strict: same adapter model, same driver version, and a network path that tolerates the extra bytes PRP appends to each frame. Hitachi Energy supports the same approach on SYS600 station computers with dual network adapters, which is why PRP shows up on almost every IEC 61850 station bus design that uses MicroSCADA.

Behind a redundancy box. The server keeps one ordinary NIC and stays a singly attached node. A RedBox in front of it does the duplication and the discard. Siemens sells the SCALANCE X204RNA for this. RUGGEDCOM, Hirschmann, and Moxa all sell equivalents.

HSR at the server is effectively always the second option. An HSR node sits in a ring and has to forward every frame it does not consume, which needs dedicated hardware rather than a driver. No mainstream SCADA platform implements an HSR stack in software. If your process bus is an HSR ring, the station computers attach through a RedBox.

That combination is the standard substation pattern: HSR rings down at the process level among merging units and relays, PRP up at the station level for the station computers and the SCADA interface, with RedBoxes joining the two.

One point that costs people time on commissioning: fault tolerance inside the chassis does not give you network redundancy. A Stratus ftServer survives a failed CPU or memory board without dropping a frame. It still needs PRP or a RedBox to survive a failed switch.

Native PRP support is stated per software version and per operating system. Confirm it against the release you are actually installing, not the marketing page.

What it costs, and how licensing works

The SCADA market splits cleanly into two camps: vendors who publish a price list, and vendors who make you ask.

Two publish. Everyone else quotes.

Ignition: full list published

Inductive Automation puts the entire catalog on its own site, with a running total as you add items. The platform license is $1,200. That is the base, and it already includes unlimited tags, unlimited clients, and connections to any supported PLC or database.

Functionality comes from modules or bundled solution suites:

Four points matter more than the numbers.

Licenses are perpetual and sold per server. One license covers one gateway, so a five-gateway architecture needs five licenses.

Redundancy costs double. Both the master and the backup gateway need a license, and the backup generally carries the same modules as the master. People forget this when budgeting.

Support plans are a percentage of retail, billed annually: 16%, 20%, or 24% depending on tier. The percentage is locked at purchase and does not fall if you later stop using a module.

Skipping the support plan is expensive later. A major version upgrade without one costs 65% of the software’s current retail price.

Ignition Edge is the cheap end for remote sites. It runs a limited module set, has no database connectivity, holds 35 days of internal history, and caps you at two concurrent client views. Devices and tags are unlimited. Bulk discounts apply.

Ignition Maker Edition is free for non-commercial use.

VTScada: price lists published as PDFs

Trihedral publishes US and Euro price lists on its own site, refreshed roughly annually.

Licensing is per seat and banded by I/O tag count, where an I/O tag is one that actually reads or writes to external equipment. Calculation-only tags do not count against the band. Tag counts cover every application running concurrently on that machine.

There are two license classes. Development Runtime allows configuration. Runtime allows operation and administration but not application development. Either can act as a primary server, a backup server, or a client.

Thin clients are priced separately by concurrent connection, in packs, and are shared across a redundant server pair.

The pitch is bundling. Historian, alarm dialer, graphic libraries, web server, telemetry drivers, and reports are in the base license rather than sold as options. When comparing against a per-module vendor, count what you would have to add on the other side.

VTScadaLIGHT is free, capped at 50 I/O and one thin client connection, with alarm notification disabled.

Everyone else quotes

No published prices, but the models are documented and worth knowing before you sit down with a salesperson.

A few notes on the ones that hurt.

Siemens counts runtime power tags, and the bands step up in large jumps. Sizing just above a band boundary is a common and avoidable mistake. Part numbers and regional list prices are visible through the Siemens ordering portal, so you can get close to a real figure without a salesperson.

AVEVA has moved between perpetual licensing and subscription credits more than once. Ask specifically what happens to your applications if you stop paying.

GE Vernova has announced new tiered pricing and licensing across the Proficy portfolio with the 2026 releases. If you are pricing iFIX or CIMPLICITY now, confirm which model you are being quoted under.

Utility platforms are not products you price from a list. They are tendered projects where the software is a line item inside an engineering contract.

The number that actually matters

Software is usually the smallest part of a SCADA project. Engineering, screens, database design, testing, and commissioning routinely cost three to ten times the license.

That changes what you should optimize for. A platform that is $10,000 cheaper but takes two extra months of engineering is not cheaper. A platform your team already knows is often the lowest total cost even at a higher list price.

Published prices are list, before any distributor or volume discount, and they change. Treat every figure here as a starting point for a quote, not a quote.

Four platforms come up constantly in SCADA searches even though they are distributed control systems: Emerson DeltaV and Ovation, Honeywell Experion PKS, Yokogawa CENTUM VP, and ABB Ability System 800xA.

The practical difference is control ownership. A DCS owns the regulatory control loops. Controllers, I/O, and the operator interface come from one vendor as an engineered system. A SCADA system supervises controllers that someone else built.

If your project is a continuous process with hundreds of PID loops and tight batch requirements, you are in DCS territory. If you are monitoring geographically spread assets over telemetry, you are in SCADA territory. Everything between those poles is a judgment call, and both vendor camps will tell you their product covers it.

Which SCADA system should you choose?

There is no single best platform. There is a best fit for a given plant, a given team, and a given protocol set. Work through it in this order.

Start with what is already there

If the plant is 90% Allen-Bradley, FactoryTalk View SE removes an entire integration layer. If it is 90% Siemens, WinCC does the same. Fighting the installed base costs more than the license difference ever saves.

This applies to people as much as hardware. A team that has maintained InTouch for twelve years will be productive in System Platform and slow in something new. That transition cost is real and it rarely appears in the comparison spreadsheet.

Then check what you have to talk to

Substation work means IEC 61850 and IEC 60870-5-104. Utility telemetry means DNP3. Plant floor means Modbus TCP and EtherNet/IP. Control center to control center means ICCP/TASE.2.

Check whether the driver is native or an add-on with its own price tag. A platform that needs a gateway in front of it to reach your field devices has just added a box, a configuration file, and a failure point to your architecture.

Then model the license at three times your point count

Per-tag licensing punishes you at every expansion. Per-server licensing does not. Point counts grow every time someone adds a feeder, a string inverter, or a metering panel. Price the system you will have in five years, not the one you are commissioning next month.

Then ask who maintains it

The best platform with no local integrator is worse than the second-best platform with three good ones in your city. Ask who else in your region runs it, and how quickly you can get someone on site.

Matching the platform to the job

What should disqualify a platform

Four things end the evaluation early, whatever the demo looked like.

It cannot reach your field devices without a gateway you did not plan for. It cannot run on the operating system your IT department will actually support. It has no integrator within a day’s travel. Or the five-year license cost exceeds the engineering cost of the whole project.

Before you sign

Run a proof of concept with your own devices and your own protocol, not the vendor’s simulator. Connect one real relay, one real PLC, or one real RTU. Read a value, write a command, and pull the network cable while it is running.

Most of the surprises that show up in month six are visible in that first afternoon.

Protocol support at a glance

The IEC 61850 column means MMS client capability, which is the vertical path between a SCADA system and the IEDs. It does not mean GOOSE. GOOSE is horizontal traffic between IEDs on the station bus, carrying trip, block, and interlock signals in a few milliseconds. That exchange is a protection function and it stays at bay and station level. A SCADA system sits above it and reads the result through MMS reporting.

Driver availability shifts between releases and editions. Confirm against the current datasheet before you commit to an architecture.

Frequently asked questions